#62 Too many integral points in the theta-direction

[Sergejs]

Hello,

When running 17k mesh with either virbating element or point source I get this error on high frequencies.:

Too many integral points in the theta-direction!
Number of integration points = 101
Maximum allowable number of integration points = 80

This is triggered by "NC_Main.cpp" on line 452 and I can not find any guidance what to do about it.

Some more details:
- the mesh contains very large triangle edges on the bottom of the shoulders.
- the number of integration points grows with frequency so it is possible to simulate this mesh if I limit max frequency to about 11000 Hz. at 11300 Hz it exceeds the Gaussean points limit.
- I included example "NC.out"

Thanks!

[Fabian Brinkmann]

Hi Sergejs,

can you post a link to the entire Project folder, please?

Best, Fabian

[Sergejs]

OK, I am sending the project folder to you directly.

[Sergejs]

The current best guess for a practical workaround:

This is likely caused by mesh surface defects (polygon folds or deep narrow depressions). OR it was caused by use of very big triangles. So to be on the safe side, avoiding use of edges longer than about 15mm is a good approach even on areas that don't need any detail.

Clearly more testing and understanding would be useful, but this is not a fatal error: if this happens, it basically points to problems in the mesh triangles.

[Sergejs]

As commented previously - this stopper is almost certainly caused by bad mesh.
If input mesh is processed by https://github.com/cg-tub/hrtf_mesh_grading I believe this issue will not happen.